Factors Governing the Protonation State of Cysteines in Proteins: An

May 22, 2013 - Binding to a metal cation (acting as a Lewis acid) causes the cysteine's pKa to drop,21 thus facilitating sulfhydryl group deprotonatio...
0 downloads 0 Views 96KB Size
Published on Web 05/15/2002

Factors Governing the Protonation State of Cysteines in Proteins: An Ab Initio/CDM Study Todor Dudev† and Carmay Lim*,†,‡ Contribution from the Institute of Biomedical Sciences, Academia Sinica, Taipei 11529, Taiwan R.O.C., and Department of Chemistry, National Tsing Hua UniVersity, Hsinchu 300, Taiwan Received November 29, 2001. Revised Manuscript Received March 5, 2002

Abstract: The detailed mechanism of metal-cysteine binding is still poorly understood. It is not clear if every metal cation can induce cysteine deprotonation, how the dielectric medium affects this process, and the extent to which other ligands from the metal’s first and second coordination shell influence cysteine ionization. It is also not clear if the zinc cation, with its positive charge reduced by charge transfer from the first two bound cysteinates, could still assist deprotonation of the next one or two cysteines in Cys3His and Cys4 zinc-finger cores. Here, we elucidate the factors governing the cysteine protonation state in metalbinding sites, in particular in Zn‚Cys4 complexes, using a combined ab initio and continuum dielectric approach. Transition metal dications such as Zn2+ and Cu2+ and trivalent cations such as Al3+ with pronounced ability to accept charge from negatively charged Cys- are predicted to induce cysteine deprotonation, but not “hard” divalent cations such as Mg2+. A high dielectric medium was found to favor cysteine deprotonation, while a low one favored the protonated state. Polarizable ligands in the metal’s first shell that can competitively donate charge to the metal cation were found to lower the efficiency of the metal-assisted cysteine deprotonation. The calculations predict that the zinc cation could assist deprotonation of all the cysteines during the folding of Cys4 zinc-finger cores and the [Zn‚(Cys-)4]2- state is likely to be preserved in the final folded conformation of the protein provided the binding site is tightly encapsulated by backbone peptide groups or lysine/arginine side chains, which stabilize the ionized cysteine core.

Introduction

One-half of the 20 natural-occurring amino acids are found to participate in metal binding in proteins. Among the metal ligands, the most common coordinating groups are the side chains of histidine, cysteine, aspartic, and glutamic acid. Although binding to a metal cation induces cysteine deprotonation under physiological conditions, it is not clear (a) if every metal cation can induce cysteine deprotonation, (b) if the zinc cation, with its positive charge reduced by charge transfer from the first two bound cysteinates, could still facilitate the deprotonation of the next one or two cysteines in Cys3His and Cys4 zinc-finger cores, (c) how other ligands from the metal’s first and second coordination shell influence cysteine ionization, and (d) how the dielectric medium affects cysteine deprotonation. Hence, in this work we have elucidated the factors governing the cysteine protonation state in metal-binding sites using a combined ab initio and continuum dielectric approach. Although emphasis has been placed on zinc-cysteine interactions, the conclusions drawn here can be applied to elucidate certain aspects of complex formation between cysteine and other transition metals of biological interest such as Cu, Ni, and Fe. The copper cation binds cysteine in azurin, pseudo-azurin, * To whom correspondence should be addressed. E-mail: carmay@ gate.sinica.edu.tw. † Institute of Biomedical Sciences. ‡ National Tsing Hua University. 10.1021/ja012620l CCC: $22.00 © 2002 American Chemical Society

amicyanin, plastocyanin, and stellacyanin,1 while nickel and iron complexes with cysteine(s) are observed in Ni-Fe hydrogenase1 and the large class of iron-sulfur proteins.2,3 Aspartates/glutamates do not appear to be very metal-specific as they coordinate to hard (Mg2+, Ca2+, Mn2+), borderline (Zn2+, Cu2+, Co2+), and soft (Cd2+, Hg2+) metal cations.1,4-7 At physiological pH (pH ≈ 7), aspartic acid and glutamic acid side chains that possess typical pKa values between 4 and 58 would be deprotonated. These side chains tend to bind directly to the metal (inner-sphere mode) if they are located in a protein cavity or crevice with a low dielectric constant, whereas they tend to bind indirectly to the metal (outer-sphere mode) on solvent-exposed protein surfaces.9 Histidine side chains exhibit greater metal selectivity than carboxylates, as they prefer borderline (Zn2+, Cu2+, Co2+, Fe2+, Ni2+) and soft (Cd2+, Hg2+) metal cations to hard ones.1,4,7 Histidines (up to three per metal cation) are often found in metalbinding sites that play a predominantly catalytic role.10-14 (1) (2) (3) (4) (5) (6) (7) (8) (9)

Rulisek, L.; Vondrasek, J. J. Inorg. Biochem. 1998, 71, 115-127. Huber, C.; Wachtershauser, G. Science 1998, 281, 670-672. Beinert, H. J. Bioinorg. Chem. 2000, 5, 2-15. Jernigan, R.; Raghunathan, G.; Bahar, I. Curr. Opin. Struct. Biol. 1994, 4, 256. Christianson, D. W.; Cox, J. D. Annu. ReV. Biochem. 1999, 68, 33-57. Dudev, T.; Cowan, J. A.; Lim, C. J. Am. Chem. Soc. 1999, 121, 76657673. Dudev, T.; Lin, Y. L.; Dudev, M.; Lim, C. 2002, in preparation. Stryer, L. Biochemistry, 4th ed.; W. H. Freeman and Co.: New York, 1995. Dudev, T.; Lim, C. J. Phys. Chem. B 2000, 104, 3692-3694. J. AM. CHEM. SOC. 2002, 124, 6759-6766

9

6759

ARTICLES

Dudev and Lim

Histidine side chains usually have pKa values close to physiological pH,8 but they are likely to be neutral upon binding to a positively charged metal ion. They can bind directly to metal dications (Zn2+, in particular) in preformed protein cavities or on solvent-exposed surfaces.9 Furthermore, theoretical studies suggest that a neutral histidine side chain, bridging between a metal cation and an acidic residue, may become deprotonated to an anionic imidazolate; that is, the metal-imidazole-carboxylate triad may isomerize to metal-imidazolate-carboxylic acid by transferring a proton from the imidazole ring to the carboxylate group.15-17 Cysteine appears to be the most enigmatic and elusive amino acid involved in metal binding. Like histidine, it prefers to bind to borderline and soft metal cations, but unlike histidine, cysteines (up to four per metal cation) are usually found in metal-binding sites that are mainly involved in protein structure stabilization as opposed to catalysis.10-14 At physiological pH, cysteine side chains with typical pKa values between 8 and 918-20 would be protonated in metal-free proteins. Binding to a metal cation (acting as a Lewis acid) causes the cysteine’s pKa to drop,21 thus facilitating sulfhydryl group deprotonation under physiological conditions. Thus, the change in the protonation state of cysteine at neutral pH is a metal-assisted process. While it is commonly accepted that the classical Zn‚Cys2His2 zinc-finger core is neutral, the protonation state of cysteines in zinc-finger cores containing either four cysteines22 or a histidine and three cysteines23 remains a subject of controversy. A consensus exists in the literature that the first two cysteines bind Zn2+ in an ionized form. However, current experimental data seem to present conflicting results on the protonation state of the third or fourth cysteine. From a theoretical viewpoint, it is also not certain if the zinc cation, with its positive charge reduced by charge transfer from the first two bound cysteinates, could still assist deprotonation of the next one or two cysteines. Experimental support for the deprotonation of only two cysteines comes from cobalt d-d spectroscopy and electrospray mass spectrometry. Garmer and Krauss24 have noticed that in Co2+liver alcohol dehydrogenase the charge transfer spectrum of the structural site consisting of four cysteines is nearly identical to that of the catalytic site consisting of two cysteines, a histidine, and a water molecule, suggesting that only two cysteines in each site are deprotonated. Using electrospray mass spectrometry, Fabris et al.25,26 have obtained accurate mass measurements of (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24) (25)

Vallee, B. L.; Auld, D. S. Biochemistry 1990, 29, 5647-5659. Christianson, D. W. AdV. Protein Chem. 1991, 42, 281-355. Coleman, J. E. Annu. ReV. Biochem. 1992, 61, 897-946. Lipscomb, W. N.; Strater, N. Chem. ReV. 1996, 96, 2375-2433. Berg, J. M.; Godwin, H. A. Annu. ReV. Biophys. Biomol. Struct. 1997, 26, 357. El Yazal, J.; Pang, Y.-P. J. Phys. Chem. B 1999, 103, 8773-8779. El Yazal, J.; Roe, R. R.; Pang, Y.-P. J. Phys. Chem. B 2000, 104, 66626667. Dudev, T.; Lim, C. J. Phys. Chem. B 2001, 105, 4446-4452. Dawson, R. M. C.; Elliot, D. C.; Elliot, W. H.; Jones, K. M. Data for Biochemical Research, 3rd ed.; Clarendon Press: Oxford, 1995. Lide, D. R., Ed. Handbook of Chemistry and Physics; CRC Press: Boca Raton, FL, 1998. Bombarda, E.; Morellet, N.; Cherradi, H.; Spiess, B.; Bouaziz, S.; Grell, E.; Roques, B. P.; Mely, Y. J. Mol. Biol. 2001, 310, 659-672. Hightower, K. E.; Huang, C.-C.; Casey, P. J.; Fierke, C. A. Biochemistry 1998, 37, 15555-15562. Petkovich, M.; Brand, N. J.; Krust, A.; Chambon, P. Nature 1987, 330, 444. Summers, M. F.; South, T. L.; Kim, B.; Hare, D. Biochemistry 1990, 29, 329. Garmer, D. R.; Krauss, M. J. Am. Chem. Soc. 1993, 115, 10247-10257. Fabris, D.; Zaia, J.; Hathout, Y.; Fenselau, C. J. Am. Chem. Soc. 1996, 118, 12242-12243.

6760 J. AM. CHEM. SOC.

9

VOL. 124, NO. 23, 2002

Cys2His2, Cys3His, and Cys4 zinc-finger peptides as well as Cys3His and Cys4 zinc proteins. They have concluded that only two of the cysteines in each binding site are ionized, while the remaining cysteine(s) appear to be protonated leading to a total charge of zero for the respective zinc complexes. On the other hand, other experimental and theoretical data support the deprotonation of all metal-bound cysteines. Maynard and Covell27 have analyzed the electrostatic screening of zincfinger cores and have found that Cys4 cores are much more positively screened by the protein than Cys3His sites, which in turn are better shielded than the classical Cys2His2 zinc-finger cores. Hence, they have concluded that Cys4 and Cys3His cores are most likely anionic in the native protein; that is, all the cysteines are deprotonated. Konrat et al.28 had earlier inferred the same conclusion from NMR analyses of the second-shell packing around Cys3His and Cys4 cores in quail cysteine-rich and glycine-rich CRP2 protein. Using potentiometric titration, NMR, and fluorescence spectroscopy, Bombarda et al.20 have concluded that in aqueous solution the three cysteines in the Cys3His motif of HIV-1 nucleocapsid protein are sequentially deprotonated in the presence of Zn2+. Ryde29 has used ab initio calculations together with X-ray data to support the deprotonation of all the cysteines in Zn‚Cys4 binding sites. In the MP2optimized [Zn‚(HS-)4]2- structure, the four Zn-S- bond distances are nearly identical (2.41 Å), whereas in the alternative structure, [Zn‚(H2S)2‚(HS-)2]0, the Zn-S0 (2.75 Å) and ZnS- (2.22 Å) distances differ by more than 0.5 Å.29 Such inequality among the Zn-S bond distances is not observed in the X-ray structure of Cys4 or Cys3His zinc-finger protein, where the Zn-S distances are found within a narrow range between 2.3 and 2.4 Å,1,7 suggesting that the zinc-bound cysteines are all deprotonated. Here, by combining ab initio and continuum dielectric calculations, we attempt to elucidate the protonation state of cysteines in Zn‚Cys4 complexes. To this end, we have systematically assessed the effect of various factors governing the protonation state of cysteine(s) in metal-binding sites, such as the dielectric medium, the nature of the metal cation, and the effect of the other first- and second-shell ligands. The protonation state of cysteine-rich zinc cores is discussed in light of the present findings. Methods Models Used. Modeling cysteine as H2S24,29-32 or CH3SH24,29,32-36 has been shown to reproduce the geometry of zinc-cysteinate cores.37 However, H2S may represent cysteine better than CH3SH in modeling reactions that affect the protonation state of cysteine since the pKa of H2S (6.9/7.038,39) is closer to that of cysteine (8.1/8.418,19) as compared to that of CH3SH (10.3/14.240,41). To verify the use of H2S as a cysteine (26) Fabris, D.; Hathout, Y.; Fenselau, C. Inorg. Chem. 1999, 38, 1322-1325. (27) Maynard, A. T.; Covell, D. G. J. Am. Chem. Soc. 2001, 123, 1047-1058. (28) Konrat, R.; Weiskirchen, R.; Bister, K.; Krautler, B. J. Am. Chem. Soc. 1998, 120, 7127-7128. (29) Ryde, U. Eur. Biophys. J. 1996, 24, 213-221. (30) Ryde, U. Int. J. Quantum Chem. 1994, 52, 1229. (31) Deerfield, D. W., II; Pedersen, L. G. J. Mol. Struct. (THEOCHEM) 1997, 419, 221-226. (32) Cini, R. J. Biomol. Struct. Dyn. 1999, 16, 1225-1237. (33) Garmer, D. R.; Gresh, N. J. Am. Chem. Soc. 1994, 116, 3556-3567. (34) Topol, I. A.; Casas-Finet, J. R.; Gussio, R.; Burt, S. K.; Erickson, J. W. J. Mol. Struct. (THEOCHEM) 1998, 423, 13-28. (35) Topol, I. A.; Nemukhin, A. V.; Chao, M.; Iyer, L. K.; Tawa, G. J.; Burt, S. K. J. Am. Chem. Soc. 2000, 122, 7087-7094. (36) Rulisek, L.; Havlas, Z. J. Am. Chem. Soc. 2000, 122, 10428-10439. (37) Dudev, T.; Lim, C. J. Phys. Chem. B 2001, 105, 10709-10714. (38) Pearson, R. G. J. Am. Chem. Soc. 1986, 108, 6109-6114.

Protonation State of Cysteines in Proteins

ARTICLES

Table 1. Calculated Solution Free Energies ∆G80 (in kcal/mol) for the Deprotonation Reaction [Zn‚W3‚LH]2+ f [Zn‚W3‚L]+ + H+ ligand LH

∆G80

pKa

H 2S CH3SH CH3CH2SH H2N(COOH)C(H)CH2SH (cysteine)

-18.8 -8.8 -10.1 -16.0

-13.8 -6.5 -7.4 -11.8

model for reactions involving a change in the cysteine protonation state, deprotonation free energies of different cysteine models (H2S, CH3SH, CH3CH2SH) bound to Zn2+ in aqueous solution were evaluated (see below) and compared with that of Zn2+-bound cysteine (H2N(COOH)C(H)CH2SH). The results in Table 1 show that the pKa of the Zn2+bound H2S (-13.8) is much closer to that of Zn2+-bound cysteine (-11.8) than the pKa values of Zn2+-bound CH3SH (-6.5) and CH3CH2SH (-7.4). Because the focus of this work is on deprotonation/ reprotonation of metal-free or metal-bound cysteine, the cysteine side chain was modeled by H2S. The histidine and lysine side chains and the backbone peptide group were modeled by imidazole, CH3NH3+, and CH3-CO-NH-CH3, respectively. As for H2S, the experimental pKas of imidazole (6.7/7.038,42), CH3NH3+ (10.642), and CH3-CONH-CH3 (15.143) are close to the respective values for histidine (6.5), lysine (10.544), and the backbone peptide group (∼1543) in aqueous solution. The preferred coordination number in zinc and copper complexes is four;1,4,7 hence their complexes were modeled as ML4 (M ) Zn2+, Cu2+; L ) H2O, H2S, HS-, imidazole). On the other hand, magnesium and aluminum prefer to be octahedrally coordinated,4,6,45 and their complexes were modeled as ML6 (M ) Mg2+, Al3+; L ) H2O, H2S, HS-). Ab Initio Calculations. In a recent study,37 we have shown that among various theoretical methods and basis sets, SVWN/6-311++G** and MP2/6-311++G**//SVWN/6-311++G** are best suited for reproducing the geometrical parameters and energetics of cysteine-like zinc complexes, respectively. Consequently, full geometry optimization for all the complexes studied in this work was carried out using the Gaussian 98 program46 at the SVWN/6-311++G** level. Vibrational frequencies were then computed at the same level of theory/basis to verify that each complex was at the minimum of its potential energy surface. No imaginary frequency was found in any of the complexes. After scaling the frequencies by an empirical factor of 0.9833,47 the zero-point energy (ZPE), thermal energy (ET), work (PV), and entropy (S) corrections were evaluated using standard statistical mechanical formulas.48 The electronic energies Eelec were then evaluated at the MP2/ 6-311++G**//SVWN/6-311++G** level (see above). The differences ∆Eelec, ∆ZPE, ∆ET, ∆PV, and ∆S between the products and reactants (39) Atkins, P. W. Physical Chemistry, 6th ed.; Oxford University Press: Oxford, 1999. (40) Howard, P. H., Meylan, W. M., Eds. Handbook of Physical Properties of Organic Chemicals; Lewis Publishers: Boca Raton, FL, 1997. (41) Irving, R. J.; Nelender, L.; Wadso, I. Acta Chem. Scand. 1964, 18, 769. (42) Smith, R. M.; Martell, A. E. Critical Stability Constants; Plenum Press: New York, 1989; Vol. 2, Suppl. 2. (43) Sigel, H.; Martin, R. B. Chem. ReV. 1982, 82, 385-426. (44) Dictionary of Organic Compounds, 6th ed.; Chapman & Hall: London, 1996. (45) Marcus, Y. Chem. ReV. 1988, 88, 1475-1498. (46) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.5; Gaussian, Inc.: Pittsburgh, PA, 1998. (47) Wong, M. W. Chem. Phys. Lett. 1996, 256, 391. (48) McQuarrie, D. A. Statistical Mechanics; Harper and Row: New York, 1976.

were employed to compute the reaction free energy at room temperature, T ) 298.15 K, according to the following:

∆G1 ) ∆Eelec + ∆ZPE + ∆ET + ∆PV - T∆S

(1)

Continuum Dielectric Calculations. The reaction free energy in a given environment characterized by a dielectric constant  ) x can be calculated according to the following thermodynamic cycle:

∆G1 is the gas-phase free energy computed using eq 1. ∆Gsolvx is the free energy for transferring a molecule in the gas phase to a continuous solvent medium characterized by a dielectric constant, x. By solving Poisson’s equation using finite difference methods49,50 to estimate ∆Gsolvx (see below), the reaction free energy in an environment modeled by dielectric constant x, ∆Gx, can be computed from:

∆Gx ) ∆G1 + ∆Gsolvx(products) - ∆Gsolvx(reactants)

(2)

The continuum dielectric calculations employed a 71 × 71 × 71 lattice with an initial grid spacing 1.0 Å, and refined with a spacing of 0.25 Å, ab initio geometries, and natural bond orbital (NBO) atomic charges.51 The low-dielectric region of the solute was defined as the region inaccessible to contact by a 1.4 Å-radius sphere rolling over the molecular surface. This region was assigned a dielectric constant of 2 (in ) 2) to account for the electronic polarizability of the solute. The molecular surface was defined by effective solute radii, which were obtained by adjusting the CHARMM (version 22)52 van der Waals radii to reproduce the experimental hydration free energies of the metal cations and ligands as well as the pKa of the model ligand compounds. Some of these solute radii have been optimized in our previous study to reproduce the pKa of H2S, the hydration free energy of Zn2+, and the reaction free energy of Zn2+ with 2,3-dimercapto-1-propanol.37 These are (in Å) RZn ) 1.40, RC ) 1.80, RS(H2S) ) 2.50, RS(HS-) ) 2.29, RO(H2O) ) 1.67, RH(H2O) ) 0.93, RH(C, H2S) ) 1.468, RH(HS-) ) 1.05. The rest of the atomic radii (RMg ) 1.50 Å, RCu ) 1.35 Å, RAl ) 1.39 Å, RO(CO) ) 1.77 Å, RN ) 1.77 Å, RH(H+) ) 0.65 Å; RH(imidazole N) ) 1.2 Å, RH(amine N) ) 1.1 Å, and RH(ammonium N) ) 0.93 Å) have been optimized to reproduce the experimental hydration free energies of the ions and neutral ligands as well as the pKa values in Table 2. Buried or partially buried metal-binding sites were characterized by an external dielectric constant out equal to 2 or 4,49,53 whereas fully solvent-exposed sites were modeled by an out equal to 80. Thus, Poisson’s equation was solved with out equal to 1, 2, 4, or 80 and in ) 2. The difference between the computed electrostatic potentials in a given dielectric medium ( ) x) and in the gas phase ( ) 1) yielded the solvation free energy ∆Gsolvx of the metal complex.

Results

Deprotonation of H2S. The calculated enthalpies and free energies of deprotonating metal-free and metal-bound H2S in different dielectric media are given in Table 3. In the gas phase, all of the reactions in Table 3 are enthalpy driven; the relative (49) (50) (51) (52)

Gilson, M. K.; Honig, B. Biopolymers 1986, 25, 2097. Lim, C.; Bashford, D.; Karplus, M. J. Phys. Chem. 1991, 95, 5610-5620. Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. ReV. 1988, 88, 899-926. Brooks, B. R.; Bruccoleri, R. E.; Olafson, B. D.; States, D. J.; Swaminathan, S.; Karplus, M. J. Comput. Chem. 1983, 4, 187-217. (53) Harvey, S. C.; Hoekstra, P. J. Phys. Chem. 1972, 76, 2987. J. AM. CHEM. SOC.

9

VOL. 124, NO. 23, 2002 6761

ARTICLES

Dudev and Lim

Table 2. Calculated Gas-Phase (∆G1, Eq 1) and Hydration (∆Gsolv80) Free Energies (in kcal/mol) and pKa Valuesa ∆Gsolv80 species

calc

expt

% diff

H+ Mg2+ Cu2+ Al3+ Im

-259.0 -450.3 -498.3 -1096.0 -10.5

-262.5a -455.5b -498.7b -1103.3b -10.2c

1.33 1.14 0.08 0.66 2.94

∆G1 d

BH f B

HS-

∆Gsolv80 d,e

(-0.7)i

CH3CONHCH3 f CH3CONCH3CH3NH3+ f CH3NH2 199.5 (205.7)m

-4.9 (-4.6)n

H2S f

pKa d

344.1 6.7 (6.9 j; 7.0k) -0.9 345.6g; 347.1h) 345.0 -10.5 (-10.0)c 14.1 (15.1)l (345.4 f;

10.1 (10.6)o

a From Gilson and Honig, 1988.60 b From Burgess, 1978.61 c From Wolfenden, 1978.62 d Values in parentheses are the experimental values. e For the deprotonation reactions, the hydration free energies of the neutral species, for which there is experimental data, are given. f From Rosenstock et al., 1977.63 g From Cumming et al., 1978.64 h From Bartmess et al., 1979.65 i From Chambers et al., 1996.66 j From Pearson, 1986.38 k From Atkins, 1999.39 l Measured for CH3CONH2.43 m From Lias et al., 1984.67 n From Cabani et al., 1981.68 o From Smith and Martell, 1989.42

contribution of the respective entropy term to the gas-phase free energy ∆G1 is small and positive. Ionizing metal-free hydrogen sulfide in the gas phase is very costly as evidenced by a large positive ∆G1 of 344 kcal/mol (Table 3). Favorable solvation of the charged reaction products, particularly hydration of the small H+ (see Table 2), substantially decreases the free energy, which drops to 9.1 kcal/mol in aqueous solution ( ) 80). This value corresponds to a pKa of 6.7, in close agreement with the experimental values of 6.9 and 7.0.38,39 The positive free energies obtained for the entire range of  imply that complete deprotonation of metal-free H2S at neutral pH is unlikely regardless of the dielectric environment. The Effect of the Metal Cation. To assess the role of the metal cation in deprotonating hydrogen sulfide, we examined the deprotonation of M‚Wi‚H2S complexes (M ) Zn2+, Cu2+, Mg2+, and Al3+; W ) water; i ) 3 or 5, see Table 3). Upon binding to a divalent metal like Zn2+, Cu2+, and Mg2+, the gasphase basicity for H2S decreases significantly (Table 3) due to stabilization of the deprotonated HS- by the metal cation (see below). As for ∆G1, the condensed-media ∆Gx values (x ) 2, 4) for metal-bound H2S are also less positive than the respective free energies for H2S. Notably, the ∆G80 for zinc- and copperbound H2S becomes negative (-19 and -30 kcal/mol, respectively) in contrast to that of metal-free H2S, but ∆G80 for magnesium-bound H2S remains positive (26 kcal/mol). This implies that when cysteine, as modeled by H2S, binds to zinc and copper on solVent-exposed protein surfaces, its deprotonation is thermodynamically favorable, whereas if it coordinates to magnesium, it is unlikely to be deprotonated regardless of the dielectric environment. Because the respective complexes of the three divalent metals have the same charge for the protonated [M‚Wi‚H2S]2+ and deprotonated [M‚Wi‚HS]+ state, their solvation free energies in a medium of dielectric constant  do not differ significantly. For example, the solvation free energy difference between the deprotonated and protonated complexes of zinc, ∆Gsolvx([Zn‚W3‚HS-]+) - ∆Gsolvx([Zn‚W3‚H2S]2+), is similar to that of magnesium and is equal to 68 kcal/mol for x ) 2, 102 kcal/ mol for x ) 4, and 135-136 kcal/mol for x ) 80. Thus, it is 6762 J. AM. CHEM. SOC.

9

VOL. 124, NO. 23, 2002

the gas-phase free energy as opposed to solvation factors that dictates the differences among the three divalent metals. The stability of the divalent metal complexes in the gas phase is governed not only by electrostatic interactions, but also by charge transfer from the ligand to the cation.33 The NBO analyses carried out for the zinc, copper, and magnesium complexes reveal that the three metal ions accept the charge transferred by their ligands to varying degrees. This is evidenced in Table 4 where the effective charge on the metal cation is anticorrelated with the amount of charge transferred from the ligands to the metal. Among the metals in Table 4, copper is the best charge acceptor with the largest CTLfM, while magnesium is the poorest. As expected, the negatively charged HS- transfers more charge to a given metal than does the neutral H2S, and is therefore stabilized by a given metal more than is H2S, as evidenced by the smaller cost of deprotonating H2S in the gas phase when it is metal bound (Table 3). The ∆CTLfM values appear to be anticorrelated with the gas-phase deprotonation free energies: the bigger the ∆CTLfM difference, the smaller the |∆G1| (Table 3). Relative to neutral H2S, the two transition metal cations (Cu2+ and Zn2+) stabilize the deprotonated HS- more than magnesium due mainly to stronger chargetransfer effects in the former as compared to the latter. Thus ∆G1 for the zinc or copper complex is less positive than that for [Mg‚W5‚H2S]2+ (by 43 or 76 kcal/mol). The effect of a trivalent cation (Al3+) on H2S ionization was also studied (Table 3). As compared to the series of divalent cations, ∆H1 and ∆G1 for the respective aluminum-assisted reaction are much less positive due to the stronger chargecharge interaction between HS- and the trivalent cation in [Al‚W5‚HS-]2+. Solvation effects are more pronounced for the aluminum species than for the respective divalent metal complexes. The solvation free energy differences between the deprotonated and protonated aluminum complexes, ∆Gsolvx([Al‚W5‚HS-]2+) - ∆Gsolvx([Al‚W5‚H2S]3+), which equals 113, 170, and 225 kcal/mol for x ) 2, 4, and 80, respectively, are significantly more positive than the corresponding values for the divalent metal complexes in Table 3 (see above). Consequently, despite the less positive gas-phase free energy, ∆G80 for the aluminum-induced H2S deprotonation is more positive than that for the zinc- and copper-assisted processes. Nevertheless, the pKa of [Al‚W5‚H2S]3+ remains negative, indicating that trivalent Al3+, like the divalent Cu2+ and Zn2+ cations, can deprotonate cysteine-like species on water-exposed surfaces. The Effect of the First Coordination Shell Ligands. The stability of Zn2+ complexes with thiolate(s) is expected to depend, among other factors, on the electrostatic properties of the neighboring ligands and their ability to competitively donate charge to the metal cation (see above). To assess the role of these factors in deprotonating the zinc-bound H2S, we examined the deprotonation of Zn‚W3-i‚Li‚H2S complexes (i ) 1, 2) in which a protein ligand L (imidazole or HS-) replaces one or two water molecules in the reference Zn‚W3‚H2S complex (see Table 5). As compared to water, imidazole is more polarizable and donates more charge to the metal cation,33,54 thus competing with H2S/HS- for Zn2+. Exchanging a water molecule in Zn‚ W3‚H2S for an imidazole decreases the positive charge on zinc, which makes the metal less effective in ionizing H2S. The ∆CTLfM drops from 0.14e in the Zn‚W3‚H2S/Zn‚W3‚(HS-) (54) Dudev, T.; Lim, C. J. Am. Chem. Soc. 2000, 122, 11146-11153.

Protonation State of Cysteines in Proteins

ARTICLES

Table 3. Calculated Deprotonation ∆H1 and ∆Gx Values (in kcal/mol) of Free and Metal-Bound H2S for Media of Different Dielectric Constant xa reaction

HS- +

∆CTLfM (e)b

∆H 1

∆G1

∆G2

∆G4

∆G80

pKa c

0.14 0.25 0.04

350.5 117.0 82.2 157.5 31.6

344.1 109.1 76.3 152.5 24.0

188.9 57.7 32.7 100.8 17.2

104.9 25.9 6.5 69.4 8.2

9.1 -18.8 -29.8 26.1 -13.9

6.7 -13.8 -21.9 19.2 -10.2

H+

H2S f [Zn‚W3‚H2S]2+ f [Zn‚W3‚HS-]+ + H+ [Cu‚W3‚H2S]2+ f [Cu‚W3‚HS-]+ + H+ [Mg‚W5‚H2S]2+ f [Mg‚W5‚HS-]+ + H+ [Al‚W5‚H2S]3+ f [Al‚W5‚HS-]2+ + H+

a All energies in kcal/mol; x ) 1 corresponds to gas-phase values, x ) 2 or 4 represents buried or partially buried metal-binding sites, whereas x ) 80 represents fully solvent-exposed sites (see Methods); W ) water. The experimental hydration free energy of the proton (∆Gsolv80(H+) ) -262.5 kcal/mol) was used in computing ∆G80, while the calculated ∆Gsolv4(H+) of -185.3 kcal/mol and ∆Gsolv2(H+) of -119.6 kcal/mol were used in evaluating ∆G4 and ∆G2, respectively. b ∆CTLfM ) CTLfM[M‚Wi‚HS-] - CTLfM[M‚Wi‚H2S], where the CTLfM values are given in Table 4. c Calculated values for  ) 80.

Table 4. NBO Charges on Metal Cations (qM) in Metal Complexes Containing Water and H2S or HSspecies

[Zn‚W3‚H2S]2+ [Zn‚W3‚HS-]+ [Cu‚W3‚H2S]2+ [Cu‚W3‚HS-]+ [Mg‚W5‚H2S]2+ [Mg‚W5‚HS-]+ b

acceptor metal orbitalsa

4s (-0.57) 4s (-0.59) 3dyz (-1.12) 3s (-0.44)

qM (e)

CTLfM (e)b

1.53 1.39 1.17 0.92 1.73 1.69

0.47 0.61 0.83 1.08 0.27 0.31

a Orbital energy (in au) for the bare metal cation is given in parentheses. Charge transferred from ligands to the metal cation.

couple to 0.09e in the imidazole-substituted complexes (Table 5). Consequently, the gas-phase deprotonation free energy of Zn‚W2‚Im‚H2S increases by ∼21 kcal/mol relative to that of Zn‚W3‚H2S (Table 5). This affects ∆Gx (x ) 2, 4, or 80), which is also more positive than the respective numbers in the reference system, although ∆G80 remains negative indicating that the deprotonation reaction, Zn‚W2‚Im‚H2S f Zn‚W2‚Im‚(HS-), is likely to occur in aqueous solution. The effect of the neighboring ligands on thiol deprotonation is even more pronounced when HS- replaces one or two water molecules in Zn‚W3‚H2S. Introducing negatiVely charged residue(s) in the metal’s first coordination shell greatly reduces the ability of Zn2+ to assist the H2S deprotonation. This is evidenced by the significant increase in the respective gas-phase free energy: ∆G1 increases from 109 kcal/mol for the reference Zn‚W3‚H2S complex to 204 and 294 kcal/mol for the monoand di-HS- substituted complexes, respectively. As found for the metal complexes in Table 3, the observed increase in ∆G1 in Table 5 is correlated with a drop in ∆CTLfM, which decreases from 0.14e for the reference system to 0.01e and 0.005e for Zn‚W2‚HS-‚H2S/Zn‚W2‚(HS-)2 and Zn‚W‚(HS-)2‚H2S/Zn‚W‚ (HS-)3, respectively (Table 5). The ∆Gx (x ) 2, 4, 80) is also more positive than the respective free energies in the reference system. However, due to the more favorable solvation of the products (especially of the proton) relative to the reactant, ∆G80 for the two HS--substituted reactions remains negative implying that the deprotonation reactions may take place in aqueous solution. Successive Deprotonation of [Zn(H2S)4]2+. To assess the role of electronic and solvation factors in determining the protonation state of cysteine-rich zinc cores (see Introduction), tetrahedral zinc complexes containing a different number of thiols (H2S) and thiolates (HS-) were examined. H2S is more polarizable than water but less polarizable than imidazole, and therefore its ability to donate charge to zinc is intermediate between water and imidazole.33 Consequently, relative to Zn‚ W3‚H2S, the gas-phase deprotonation free energy of Zn(H2S)4,

in which three thiols have replaced three water molecules, increases slightly (by ∼4 kcal/mol, Table 6). However, solvation effects favor the deprotonated state, and hence ∆Gx (x ) 2 and 4) is less positive, while ∆G80 is more negative (by ∼9 kcal/ mol) than the corresponding numbers for Zn‚W3‚H2S (compare Table 6 with Table 5). Because of the presence of a negatively charged ligand in the first coordination shell of the metal (see above), ∆G1 for the second deprotonation reaction increases to 205 kcal/mol even though the resulting product [Zn‚(H2S)2‚(HS-)2]0 has a net charge of zero. Consequently, ∆G80 for the second deprotonation reaction is less favorable (-12 kcal/mol) than that for the first reaction (-28 kcal/mol), but its sign remains negative. However, the third consecutive deprotonation reaction, that is, [Zn‚(H2S)2‚(HS-)2]0 f [Zn‚H2S‚(HS-)3]- + H+, could not be modeled since a stationary point for [Zn‚H2S‚(HS-)3]- could not be located; the initial tetrahedral complex, after several steps of optimization, isomerized to a trigonal semiplanar structure with H2S transferred to the second coordination shell. These results are in accord with previous theoretical calculations,29 and the same result was found when H2S was replaced by CH3SH. Therefore, a two-proton-release process [Zn‚(H2S)2‚(HS-)2]0 f [Zn‚(HS-)4]2- + 2H+ was considered. As expected, deprotonating neutral [Zn‚(H2S)2‚(HS-)2]0 to yield a dianionic metal complex and two protons in the gas phase increases ∆G1 substantially (to 678 kcal/mol, Table 6) relative to the first two reactions. However, the charged products are much better solvated than the neutral reactant; thus solvation effects favor ionization. The interplay between solute electronic and solvation effects results in a negative ∆G80 (-13 kcal/mol), implying that the last reaction, as for the first two, is thermodynamically favorable in aqueous solution. The results presented in Table 6 suggest that the tetrathiol zinc complex could be fully deprotonated to the corresponding tetrathiolate species if the process occurs on water-exposed surfaces ( ) 80). Low dielectric media ( ) 2-4) characteristic of preformed solvent-inaccessible protein cavities do not seem to favor ionization of cysteine-like species, as evidenced by the large positive ∆G2 and ∆G4 in Table 6. The Effect of the Second Coordination Shell Ligands. The amide backbone and the lysine and arginine side chains, which are possible proton donors, have been found to encapsulate cysteine-rich cores in zinc fingers.27 Hence, we have explored the possibility of reprotonating [Zn‚(HS-)4]2- by a second-shell amide backbone (modeled by CH3CONHCH3) or a lysine side chain (modeled by CH3NH3+) by studying {[Zn‚ (HS-)4]‚(CH3CONHCH3)2}2- and {[Zn‚(HS-)4]‚(CH3NH3+)2}0 complexes in which N-methylacetamide and methylammonium J. AM. CHEM. SOC.

9

VOL. 124, NO. 23, 2002 6763

ARTICLES

Dudev and Lim

Table 5. Calculated Deprotonation ∆H1 and ∆Gx Values (in kcal/mol) of Zn2+-Bound H2S Complexes for Media of Various Dielectric Constant xa reaction

‚HS-]+

H+

[Zn‚W3‚H2 f [Zn‚W3 + [Zn‚W2‚Im‚H2S]2+ f [Zn‚W2‚Im‚HS-]+ + H+ [Zn‚W2‚HS-‚H2S]+ f [Zn‚W2‚(HS-)2]0 + H+ [Zn‚W‚(HS-)2‚H2S]0 f [Zn‚W‚(HS-)3]- + H+ S]2+

∆CTLfZn (e)b

∆H 1

∆G1

∆G2

∆G4

∆G80

pKa ( ) 80)

0.14 0.09 0.01 0.005

117.0 135.6 209.5 300.4

109.1 130.3 204.4 293.7

57.7 72.8 110.6 155.1

25.9 38.2 57.9 80.9

-18.8 -9.8 -7.2 -2.7

-13.8 -7.2 -5.3 -2.0

a See footnotes to Table 3. b Difference in charge transfer from ligand L to metal M: ∆CT LfZn ) CTLfZn(Zn‚L3‚HS ) - CTLfZn(Zn‚L3‚H2S) (L ) water, imidazole, HS-).

Table 6. Calculated Deprotonation ∆H1 and ∆Gx Values (in kcal/mol) of H2S in Thiol/Thiolate Zinc Complexes for Media of Different Dielectric Constant xa reaction

∆H 1

∆G1

∆G2

∆G4

∆G80

pKa ( ) 80)

f [Zn‚(H2S)3‚HS-]+ + H+ [Zn‚(H2S)3‚HS-]+ f [Zn‚(H2S)2‚(HS-)2]0 + H+ [Zn‚(H2S)2‚(HS-)2]0 f [Zn‚(HS-)4]2- + 2H+

121.8 210.3 692.6

113.1 204.7 678.1

53.4 107.2 354.6

18.5 53.0 181.2

-27.7 -12.2 -13.0

-20.4 -9.0

[Zn‚(H2S)4]2+

a

See footnotes to Table 3.

complex I (by 56 kcal/mol for {[Zn‚(HS-)4]‚(CH3-CO-NHCH3)2}2- and 15 kcal/mol for {[Zn‚(HS-)4]‚(CH3NH3+)2}0). Thus, proton transfer from the second-shell ligand to the firstshell cysteinate seems unlikely. Although the second-shell peptide backbone groups and lysine side chains do not seem to play a role in reprotonating the firstshell cysteinates, they can contribute to the stability of the cysteine-rich cores. To assess the extent to which the secondshell peptide backbone and lysine side chain can stabilize the zinc-finger cysteine-rich core, we evaluated the formation free energy (∆Gform) of the following reactions: Figure 1. Ball-and-stick diagram of the fully optimized structure of the {[Zn‚(HS-)4]‚(CH3CONHCH3)2}2- complex.

[Zn‚(HS-)4]2- + 2(CH3CONHCH3) f {[Zn‚(HS-)4]‚(CH3CONHCH3)2}2- (3) [Zn‚(HS-)4]2- + 2(CH3NH3+) f {[Zn‚(HS-)4]‚(CH3NH3+)2}0 (4)

Figure 2. Ball-and-stick diagram of the fully optimized structure of the {[Zn‚(HS-)4]‚(CH3NH3+)2}0 complex.

were placed in the second coordination layer (see Figures 1 and 2). Two extreme cases were considered. In the first case, the amide/ammonium proton was transferred to the first-shell thiolate sulfur by fixing RS-H to 1.35 Å and RH‚‚‚N to 2.0 Å; this complex is referred to as complex I. In the second case, the proton was restrained to the second-shell amide/ammonium nitrogen by fixing RN-H to 1.03 Å and RH‚‚‚S- to 2.32 Å; this complex is referred to as complex II. Note that the N-H distance of 1.03 Å and S-H distance of 1.35 Å are typical equilibrium SVWN/6-311++G** bond distances obtained for the respective ligands. In both complexes I and II, the remaining internal coordinates were optimized. The resulting electronic energies show that complex II is more stable than the respective 6764 J. AM. CHEM. SOC.

9

VOL. 124, NO. 23, 2002

The fully optimized structures of {[Zn‚(HS-)4]‚(CH3CONHCH3)2}2- and {[Zn‚(HS-)4]‚(CH3NH3+)2}0 are shown in Figures 1 and 2, respectively. Not surprisingly, the calculated ∆Gform for reaction 4 (-286 kcal/mol) is much more negative than that for reaction 3 (-28 kcal/mol) due to the favorable chargecharge interactions between the negatively charged [Zn‚(HS-)4]2and CH3NH3+ in reaction 4. Nevertheless, the negative ∆Gform obtained for both reactions suggests that both backbone and lysine packing contribute significantly to the [Zn‚(HS-)4]2- core stability. Discussion

The Interplay between Solute Electronic and Solvent Effects in Determining the Protonation State of Cysteinelike Molecules. Both solute electronic and solvent effects contribute to the protonation/deprotonation of the thiolate/thiol molecules, and the net outcome for each reaction is a delicate balance between these two effects. Solute electronic effects strongly disfavor deprotonating metal-free H2S in the gas phase, whereas solvation (especially hydration) of the charged products, HS- and H+, favors the ionization process. Solvation effects, however, cannot effectively compensate for the large ∆G1, and the free energies for deprotonating H2S remain positive for  ranging from 2 to 80 (Table 3). Binding to a metal cation

Protonation State of Cysteines in Proteins

significantly decreases the gas-phase deprotonation free energy by up to 14 times (Table 3). Solvation effects in this case, however, are not as favorable as those for H2S deprotonation since the charged, metal-bound reactants [M‚Wi‚H2S]2+ (M ) Zn, Mg, Cu, Al; i ) 3 or 5) are better solvated than neutral, metal-free H2S. Furthermore, the favorable solvation effects are not very pronounced in low dielectric media ( ) 2-4). Therefore, solute electronic factors dominate, and the ∆G2 and ∆G4 for the deprotonation reactions in Tables 3, 5, and 6 generally adopt the same positive sign as the respective ∆G1. In aqueous solution, however, the favorable hydration free energy of the small proton (-262.5 kcal/mol) may in some cases reverse the gas-phase trends. The interplay between solute electronic and solvent effects in determining the ionization state of H2S is illustrated by Al3+bound H2S. Solute electronic effects (specifically charge-charge electrostatic interactions) cause Al3+ to exhibit the highest ∆G1 drop among the metals studied, but solvation effects make it less efficient in inducing H2S deprotonation than Zn2+ and Cu2+ in aqueous solution (Table 3). Another interesting example is the stepwise deprotonation of [Zn‚(H2S)4]2+. Because of the charge transfer from thiolate ligand(s), which decreases the Lewis acidity of Zn2+, each successive deprotonation step is characterized by a higher positiVe ∆G1 than its previous one (Table 6). Yet because of better solvation of the products, especially H+, as compared to the corresponding reactant, each successive deprotonation step is characterized by more favorable solvation effects, which eventually outweigh ∆G1 at  ) 80. Hence, all the deprotonation reactions in Table 6 are thermodynamically favorable in aqueous solution. Factors Governing Deprotonation/Reprotonation Processes. The metal cation plays a critical role in governing the protonation state of the cysteine-like species. By stabilizing the thiolate form, it facilitates proton release especially in highdielectric media. Zn2+, Cu2+, and Al3+ are predicted to successfully assist thiol deprotonation on solvent-exposed surfaces (Table 3). Not any metal cation, however, can induce deprotonation. Metals, like Mg2+, with low affinity toward sulfur-containing ligands and consequently limited ability to stabilize the respective thiolate complexes appear incapable of inducing H2S deprotonation even in aqueous solution. The dielectric medium also plays a crucial role in governing the protonation state of the cysteine-like species. The calculations suggest that deprotonation of zinc-bound H2S generally occurs on solvent-exposed surfaces where hydration of the products, especially of H+, favors the ionization process (Tables 5 and 6). In contrast, deprotonation of zinc-bound H2S is unlikely to take place in the gas phase or in low dielectric media ( e 4), which favors protonation of HS-, provided an appropriate proton donor is available nearby. The electrostatic properties of other ligands bound to the metal can also affect the ionization of H2S. Highly polarizable, neutral ligands (like imidazole) or negatively charged species (like HS-) can competitively transfer charge to the metal cation, thus decreasing its Lewis acidity. This reduces the efficiency of the metal in assisting thiol ionization and results in deprotonation free energies that are more positive than the respective numbers for deprotonation of [Zn‚W3‚H2S]2+ (see Table 5). Biological Implications. The calculations predict that Zn2+ and Cu2+ can successfully aid thiol deprotonation. In fact, among

ARTICLES

all divalent metals, zinc and copper are most frequently found bound to one or more cysteines in protein-binding sites.1,4,7,10 On the other hand, the calculations predict that Mg2+ cannot induce proton release from cysteine-like species. This may explain why (to the best of our knowledge) no magnesiumbinding sites have been found to contain cysteine in the first coordination shell of magnesium.6,55,56 The calculations also predict that trivalent metals such as Al3+ may successfully assist cysteine deprotonation. Aluminum is not a natural cofactor in proteins, and information about its binding properties is scarce. There are only three X-ray structures of aluminum-binding sites deposited in the Protein Data Bank.57 These all are aluminumcomplexed D-xylose isomerases where the natural cofactor Mg2+, which is bound to Asp, Glu, and His, has been replaced by Al3+. Thus it would be interesting to carry out experiments to verify if trivalent metals such as Al3+ could assist cysteine deprotonation. The results obtained here suggest that all cysteines in Zn‚ Cys4 structural binding sites, typical of steroid receptor-type zinc fingers,22 may become deprotonated on solVent-exposed surfaces. This may happen in the course of zinc-finger folding, which is known to be a metal-induced process whose initial steps involve the binding of zinc to water-accessible cysteines in the unfolded protein.14,58,59 In the final folded structure, however, the cysteine-rich cores are usually found to be deeply buried in the protein.7 The calculations predict that under such conditions (corresponding to a low dielectric constant), fully ionized Zn‚(Cys-)4 cores can only be preserved if they are shielded from possible proton donors. The calculations also predict that backbone peptide groups or lysine side chains in the second shell energetically stabilize tetracysteinate zinc cores, but are unlikely to reprotonate any of the Cys-. These predictions are consistent with the finding that in zinc fingers the cysteine-rich cores are surrounded by an elaborate network of backbone:core NH‚‚‚S- hydrogen bonds, which stabilize the zinc-cysteinate structure.27 Salt bridges formed between secondshell lysine or arginine side chains with first-shell cysteinate have been found to additionally stabilize the binding sites in some nuclear hormone receptors, Cys- and Gly-rich proteins, polymerases, and ring-finger proteins.27,28 Taken together, these results seem to indicate that fully deprotonated Zn‚(Cys-)4 cores are likely to exist in structural binding sites provided that they are tightly encapsulated by a second coordination shell of backbone peptide groups or Lys/Arg side chains. (55) Black, C. B.; Huang, H. W.; Cowan, J. A. Coord. Chem. ReV. 1994, 135/ 136, 165. (56) Cowan, J. A. Biological Chemistry of Magnesium; VCH: New York, 1995. (57) Abola, E. E.; Sussman, J. L.; Prilusky, J.; Manning, N. O. Protein Data Bank ArchiVes of Three-Dimensional Macromolecular Structures; Academic Press: San Diego, 1997; Vol. 277. (58) Miura, T.; Satoh, T.; Takeuchi, H. Biochim. Biophys. Acta 1998, 1384, 171. (59) Krizek, B. A.; Amann, B. T.; Kilfoil, V. J.; Merkle, D. L.; Berg, J. M. J. Am. Chem. Soc. 1991, 113, 4518-4523. (60) Gilson, M. K.; Honig, B. Proteins: Struct., Funct., Genet. 1988, 4, 7-18. (61) Burgess, M. A. Metal Ions in Solution; Ellis Horwood: Chichester, England, 1978. (62) Wolfenden, R. Biochemistry 1978, 17, 201. (63) Rosenstock, H. M.; Draxl, K.; Steiner, B. W.; Herron, J. T. J. Phys. Chem. Ref. Data 1977, 6, 736. (64) Cumming, J. B.; Kebarle, P. Can. J. Chem. 1978, 56, 1. (65) Bartmess, J. E.; McIver, R. T., Jr. Gas-Phase Ion Chemistry; Academic Press: New York, 1979; Vol. 2. (66) Chambers, C. C.; Hawkins, G. D.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. 1996, 100, 16385-16398. (67) Lias, S. G.; Liebman, J. F.; Levin, D. J. Phys. Chem. Ref. Data 1984, 13. (68) Cabani, S.; Gianni, P.; Mollica, V.; Lepori, L. J. Solution Chem. 1981, 10, 563. J. AM. CHEM. SOC.

9

VOL. 124, NO. 23, 2002 6765

ARTICLES

Dudev and Lim

Conclusions

The calculations also suggest that cysteine-rich zinc cores in proteins may be fully deprotonated if they are solvent exposed. The same protonation state [Zn‚(Cys-)4]2- is likely to be preserved in the final folded conformation of the protein provided the binding site is tightly encapsulated by peptide backbone groups or Lys/Arg side chains, which energetically stabilize the tetracysteinate core structure. However, in circumstances where the tight packing is compromised (by heat or chemical reagents, for example) and an appropriate proton donor is available nearby, reprotonation of first-shell cysteinate(s) in a low dielectric medium could not be ruled out.

By using a combination of ab initio and continuum dielectric calculations, four major factors have been found to govern the protonation state of metal-bound cysteine(s) in proteins: 1. The Dielectric Medium. A high dielectric medium favors cysteine deprotonation, while a low one favors the protonated state. 2. Properties of the Metal Cation. Transition metal dications such as Zn2+ and Cu2+ with pronounced ability to accept charge from negatively charged Cys- induce cysteine deprotonation. Trivalent cations, although rarely present as natural cofactors in proteins, are predicted to also successfully assist cysteine ionization. “Hard” divalent cations, such as Mg2+, are unlikely to induce cysteine deprotonation. 3. Properties of the First-Shell Ligands. Polarizable ligands in the metal’s first shell that can competitively donate charge to the metal cation lower the efficiency of the metal-assisted cysteine deprotonation. 4. Properties of the Second-Shell Ligands. Peptide backbone groups or lysine/arginine side chains in the metal’s second shell can form NH‚‚‚S- hydrogen bonds to stabilize a negatively charged zinc-cysteinate structure.

6766 J. AM. CHEM. SOC.

9

VOL. 124, NO. 23, 2002

Acknowledgment. We are grateful to D. Bashford, M. Sommer, and M. Karplus for the program to solve the Poisson equation. This work was supported by the National Science Council, Taiwan (NSC contract #90-2113-M-001), the Institute of Biomedical Sciences, and the National Center for HighPerformance Computing, Taiwan. JA012620L